Recombinant Silk-Elastinlike Protein Polymer Displays Elasticity

3028

Biomacromolecules 2009, 10, 3028–3036

Recombinant Silk-Elastinlike Protein Polymer Displays Elasticity Comparable to Elastin Weibing Teng,† Joseph Cappello,‡ and Xiaoyi Wu*,†,§ Department of Aerospace and Mechanical Engineering, Biomedical Engineering IDP and Bio5 Institute, University of Arizona, Tucson, Arizona 85721, and Protein Polymer Technologies, Inc., San Diego, California 92121 Received June 9, 2009; Revised Manuscript Received August 25, 2009

We evaluated the mechanical properties of the genetically engineered, recombinant silk-elastinlike protein copolymer, SELP-47K. In tensile stress-strain analysis, methanol-treated non-cross-linked SELP-47K films exceeded the properties of native aortic elastin, attaining an ultimate tensile strength of 2.5 ( 0.4 MPa, an elastic modulus of 1.7 ( 0.4 MPa, an extensibility of 190 ( 60%, and a resilience of 86 ( 4% after 10 cycles of mechanical preconditioning. Stress-relaxation and creep analysis showed that films substantially maintained their elastic properties under sustained deformation. Chemical cross-linking of SELP-47K films doubled the elastic modulus and ultimate tensile strength and enhanced the extensibility and resilience. The underlying conformational and microstructural features of the films were examined. Raman spectroscopy revealed that the silklike blocks of SELP-47K existed in antiparallel β-sheet crystals in the films, likely responsible for the robust physical crosslinks. Scanning electron microscopy (SEM) revealed that the various processing treatments and the mechanical deformation of the films induced changes in their surface microstructure consistent with the coagulation and alignment of polymer chains. These results demonstrate that films with excellent elasticity, comparable to native aortic elastin, are obtainable from SELP-47K, a protein copolymer combining both silk- and elastin-derived sequences in a single polymer chain.

Introduction Elastin that constitutes 30-50% by weight of the aorta and major arterial vessels provides elasticity to blood vessels and other soft connective tissue.1 The formation of elastin in vivo involves a complex, multistep process, including the biosynthesis,2 secretion,2 delivery,3 alignment,3 coacervation,4 and cross-linking1 of the soluble polypeptide precursor of elastin, tropoelastin. In particular, the cross-linking of elastin is achieved through the enzymatic oxidation of lysine residues of adjacent tropoelastin chains. Once cross-linked, the soluble tropoelastin forms highly insoluble elastin. Because of its extreme insolubility, structural analysis of elastin has largely been limited to soluble tropoelastin5,6 and soluble segments of elastin.7 Further, this insolubility has mostly precluded the use of native elastin in the fabrication of useful materials. For materials use in the engineering of soft connective tissues, therefore, it remains a tremendous challenge to mimic the key mechanical properties of native tissues provided by the excellent elasticity of insoluble elastin. Besides structural support, accumulating evidence has also uncovered other important biological roles of elastin in healthy tissues, such as the regulation of arterial morphogenesis.8 Thus, there is a pressing need to engineer materials with the elasticity of elastin and the capability of elastin-like biological functions for a variety of biomedical applications. Genetic engineering technology has enabled the synthesis of elastin analogs with high levels of control of their primary structure. For instance, elastin-like protein (ELP) polymers composed of repeated pentapeptide sequences I/VPGXG (one* To whom correspondence should be addressed. Tel.: 1-00-520-6265854. Fax: 1-00-520-621-8198. E-mail: [email protected]. † Department of Aerospace and Mechanical Engineering. ‡ Protein Polymer Technologies. § Biomedical Engineering IDP and Bio5 Institute.

letter amino acid abbreviation, with X specifying the position where the amino acid substitutions may be made) display interesting physical behavior such as high deformability and inverse temperature transition (ITT) or coacervation.9-12 The transition temperature of an ELP polymer can be altered by substituting the amino acid occupying the X position of the pentapeptide sequence. For materials applications, ELP polymers are often chemically cross-linked in order to attain sufficient stability and mechanical strength for uses such as tissue engineering. To avoid the addition or removal of reagents or unreacted intermediates needed for chemical cross-linking, ELP polymers that undergo physical cross-linking have been pursued. A series of recombinant ELP triblock copolymers were produced that would undergo coacervation of the end blocks in aqueous solution at ambient temperatures.12-15 Furthermore, substitution of an alanine for the glycine amino acid in the third position of the pentapeptide sequence VPGXG changed the mechanical response of the end blocks from elastic to plastic deformation. The plastic, “hard” phases formed during the coacervation of the end blocks functioned as physical cross-links for the ELP copolymers. However, our previous study showed that the mechanical resilience of the physically cross-linked ELP copolymers was far inferior to that of native elastin.16 It was postulated that the plastic deformation and mechanical failure of the physical crosslinks were responsible for the poor resilience. Our recent study demonstrated that redesigning the ELP copolymers with substantially enlarged end block segments improved their resilience and load-bearing capacity, but is still not comparable to native elastin.15 An alternative physical cross-linking strategy of ELP copolymers was explored in this study. Specifically, a silk-

10.1021/bm900651g CCC: $40.75  2009 American Chemical Society Published on Web 09/30/2009

Elasticity of Silk-Elastinlike Protein Polymer

elastinlike protein (SELP) copolymer, SELP-47K, with a monomer structure of (S)4(E)4(EK)(E)3, in which S is the silklike sequence GAGAGS, E is the elastinlike sequence VPGVG, and EK is the penta-peptide sequence VPGKG, was evaluated.17,18 The monomer is repeated 13 times in each polymer chain. This design provides a repeating interdispersion of “hard” blocks and “soft” blocks in the same polymer chain enabling each chain to form multiple physical cross-links with multiple neighboring chains. Notably, the silklike blocks of SELP-47K may be capable of crystallizing to form mechanically robust physical cross-links between the elastin-mimetic sequences, which, in turn, decrease the overall crystallinity of the polymer thus enhancing its solubility and processability. SELP-47K has been fabricated into a variety of useful structures for possible biomedical applications, such as microdiameter fibers and patterned microstructures.19,20 The focus of this study was to analyze the mechanical properties of SELP47K films, including their resilience and time-dependent viscoelastic deformation. The capacity of the crystallized silklike blocks to stabilize the SELP-47K structure and to function as physical cross-links was evaluated. In addition, lysine residues present in the SELP-47K design also allow chemical crosslinking of the elastinlike blocks and possible enhancement of the mechanical properties of SELP-47K. Accordingly, the effects on the mechanical properties of SELP-47K films after glutaraldehyde cross-linking were examined.

Biomacromolecules, Vol. 10, No. 11, 2009

3029

Figure 1. SDS-PAGE images of SELP-47K. The unit of molecular weight is kilo-dalton (KD).

Materials and Methods Sample Preparation. Frozen SELP-47K aqueous solutions at a concentration of 13% (w/w) were generously provided by Protein Polymer Technologies, Inc. (PPTI, San Diego, CA). The complete amino acid sequence of SELP-47K was previously reported elsewhere.17 Protein purity and molecular weight were analyzed using sodium dodecyl sulfate polyacrylamid gel electrophoresis (SDS-PAGE) and matrix-assisted laser desorption/ionization-time-of-flight (MALDITOF) mass spectrometry. Briefly, lyophilized protein was dissolved in DI water at a concentration of 0.1% (w/v). A total of 10 µL of protein aqueous solution was eluted in 100 µL of Laemmli sample buffer containing 5 µL of β-mercaptoethanol by boiling for 6 min, electrophoresed through 7.5% precast polyacrylamid gel (Bio-Rad) at 200 V for 29 min, and then stained with Bio-Safe Coomassie Stain after being rinsed in DI water three times. All the reagents used for SDS-PAGE analysis were obtained from Bio-Rad. The SDS-PAGE images of SELP47K showed no detectable impurity in the protein (Figure 1). However, a molecular weight of 75-80 KD determined from the SDS-PAGE analysis was much higher than the theoretical calculation of 69814 for the same protein.17 MALDI-TOF mass spectrometry was then used to more accurately measure the molecular weight of SELP-47K. MALDI-TOF mass spectroscopy was performed on a Bruker ReflexIII mass spectrometer (Fremont, CA), equipped with a N2 laser (337 nm). Lyophilized SELP-47K powder was dissolved in DI water, and 3,5-dimethoxy-4-hydroxycinnamic acid (SA) was used as matrix for MALDI-TOF mass spectroscopy. The mass spectrum of SELP-47K was relatively weak, presumably due to its high molecular weight (Figure 2). A molecular ion peak at 69699 was close to the theoretical molecular weight of SELP-47K (i.e., 69,814), assumedly corresponding to SELP-47K molecules with one charge. A discrepancy of 115 Da between the actual and theoretical molecular weights of SELP-47K is less than the half-bandwidth of the molecular ion peak. It was also anticipated that SELP-47K molecules with multiple charges displayed molecular ion peaks at 34849 () 69699/2, Z ) 2); 23233 () 69699/3, Z ) 3); 17424 () 69699/4, Z ) 4), and so on. Observed were two strong peaks at 34859 and 23195 and a weak peak at 17364. The observed peaks at low mass-to-charge ratios were consistent with the anticipated peaks, suggesting the obtained MALDI-TOF mass spectrum is the protein fingerprinting of SELP-47K.

Figure 2. MALDI-TOF mass spectrum of SELP-47K.

Lyophilized protein was dissolved in deionized water at a concentration of 20% (w/w) at room temperature. The protein solution was mixed thoroughly by Vortex, and air bubbles were removed by centrifugation. The protein solution was then poured into polypropylene casting molds and solvent evaporation was performed at room temperature for 24 h under ambient conditions. After complete solvent evaporation, SELP47K films, which were denoted as nontreated films, were gently peeled off the casting molds for analysis or further chemical treatments. Chemical Treatment. SELP-47K cast films were treated with 99.9% methanol (MeOH, Fisher Scientific) for 4 h and air-dried before further processing or mechanical analysis. After air-drying, some MeOH-treated films were cross-linked with 10 mL of glutaraldehyde (GTA, Mallinckrodt Baker) solutions at a concentration of 1% (w/v) in phosphate buffer at pH 7.4 for 24 h at room temperature, following a procedure used by Bigi et al.21 The methanol-treated and then glutaraldehyde-cross-linked films were denoted as cross-linked films. Cross-linked films were extensively rinsed using deionized water prior to mechanical and structural analysis. Mechanical Characterization. Methanol-treated non-cross-linked and cross-linked SELP-47K films were cut into rectangular samples with dimensions of 3 × 5 mm or 15 × 2 mm for mechanical analyses. The gauge length (the sample length between two clamps) for the large samples was approximately 10 mm, while that of the small samples was between 5.5 and 6.0 mm. Depending on the sample gauge length, the maximum travel distance of the drive shaft of the dynamic mechanical analyzer (DMA) was 20-24 mm. Small samples that could be broken using the current setup were used for measurement of the ultimate tensile strength and deformability of SELP-47K films. Large samples were used to more accurately analyze the mechanical behavior

3030


Teng et al. fluorescent effects, the Raman spectrum of methanol-treated glutaraldehyde-cross-linked films was not obtained. Scanning Electron Microscopy. The surface morphologies of three types of SELP-47K films, including nontreated, methanol-treated noncross-linked, and methanol-treated glutaraldehyde-cross-linked films, were examined using a Hitachi-S4800 scanning electron microscope (SEM). Typically, film samples were vacuum-dried in a desiccator overnight and then coated with platinum for 5-20 s using a sputter coater. Films were also examined after preconditioning and after being stretched to failure to evaluate their changes in microstructure.

Results and Discussion Figure 3. Representative stress-strain analysis of methanol-treated non-cross-linked and cross-linked SELP-47K films. Small stress drops in cross-linked SELP-47K films (noted by arrows) were due to the minimal time lapse between successive steps.

of SELP-47K films at small and medium deformation up to 50% strain. Samples were hydrated in 1× phosphate-buffered saline (PBS) at 37 °C, which contained 0.2 mg/mL NaN3 to prevent biological contamination. Hydrated film thickness was typically 0.2 mm, as measured by optical microscopy. Mechanical characterization of SELP-47K samples was performed using a PerkinElmer diamond DMA, and samples were immersed in a jacketed beaker filled with 1X PBS at 37 °C. Both methanol-treated non-cross-linked and cross-linked films were evaluated by several mechanical testing protocols including the following: (i) Uniaxial tensile failure. Three to five replicate small samples of each type of film were monotonically extended to failure so the deformability and ultimate tensile strength were obtained. Displacement was applied at a fixed rate of 0.25 mm/min. The maximum displacement within each loading step was 5 mm due to a limitation of the instrument, so multistep tension experiments were programmed for the failure study. Loading steps were successively initiated with a time lapse between steps of only a few seconds, inducing negligible stress relaxation in the tensile stress-strain analyses of SELP-47K films (see the small stress drops marked by arrows in Figure 3). The Young’s modulus of a sample was determined as the slope of the stress-strain curve in the first 20% deformation region. (ii) Mechanical preconditioning. A total of 10-12 samples of each type of film were cyclically stretched to approximately 50% strain (i.e., 5 mm displacement with a gauge length of approximately 10 mm) for 10 cycles with an off-loading period of 5 min between cycles. Displacement of the cross-head was applied during loading and unloading at a fixed rate of 1 mm/min. Resilience was calculated from the loading and unloading curves at a fixed cycle and then averaged across replicate samples.

%resilience ) 100 × area under unloading curve/area under loading curve

(1)

(iii) Cyclic loading behavior. The loading and unloading behaviors of preconditioned samples were examined under different cyclical deformations (e.g., 10, 20, 30, and 40% strains). Displacement was applied during loading and unloading at a constant speed of 0.5 mm/ min. (iv) Creep and stress relaxation. Three to five replicate samples were prepared for creep analysis. Constant stresses were applied for time periods of up to 15 h. Three to five replicate samples were stretched to 20% strain at a constant rate of 0.5 mm/min, and the stress over time was recorded for 15 h. Raman Spectroscopy. Raman spectra of SELP-47K films were recorded on a Thermo Nicolet Almega microRaman system (Thermo Scientific). A solid-state laser with the wavelength of 532 nm was used as the excitation source. Three types of SELP-47K films, including the nontreated films, methanol-treated noncross-linked films, and methanol-treated noncross-linked films after 10 cycles of mechanically preconditioning, were analyzed using Raman spectroscopy. Due to

SELP-47K Films Displayed High Deformability. Methanoltreated non-cross-linked SELP-47K films displayed enhanced mechanical strength as compared to nontreated films, which were unstable in 1× PBS. Fully hydrated, methanol-treated noncross-linked SELP-47K films possessed strain at failure of 190 ( 60% and ultimate tensile strength of 2.5 ( 0.4 MPa (Figure 3, n ) 3). They displayed nearly linear deformation response to external applied forces, with an estimated average Young’s modulus of 1.3 MPa. After glutaraldehyde cross-linking, tensile stress-strain analysis measured dramatic increases in the deformability (245 ( 58%) and ultimate tensile strength (5.4 ( 1.1 MPa) of cross-linked films as compared to non-crosslinked films (Figure 3, n ) 3). It is worthy of comparing the mechanical behavior of methanol-treated non-cross-linked SELP47K films to those of native silks and elastin. When analyzed in the dry state, Gosline et al. reported that both the major ampullate (MA) silk of spiders and silkworm silk showed Young’s modulus of 7-10 GPa and maximum deformability of 18-27%.22,23 Although hydration reduced the Young’s modulus of MA silk by 100-fold, Shao and Vollrath demonstrated that its ultimate tensile strength remained around 600 MPa.24 In contrast, the mechanical properties of methanoltreated non-cross-linked SELP-47K films are very comparable to those of native elastin from bovine ligament, which displayed a Young’s modulus of 1.1 MPa, ultimate tensile strength of 2 MPa, and deformability of 150%.22,25 In our model, the methanol-treated non-cross-linked SELP47K films form a three-phase structure: a semirigid phase formed by the crystallized silklike blocks, a flexible phase comprised of the hydrated elastinlike blocks, and mixed phase formed by the interpenetrating crystallized silklike and hydrated elastinlike blocks. Under mechanical force, the flexible elastinlike blocks deform first. A further increase in mechanical force will lead to the deformation and disruption of the mixed phase.26,27 Consequently, irreversible deformation and subsequently poor resilience of SELP-47K films may be observed. However, the disruption of the mixed phase also facilitates separation of the crystallized silklike and hydrated elastinlike phases. As a result, a sample may appear to be more resilient, if subsequent deformation does not exceed the previous deformation. Thus, it is anticipated that mechanical preconditioning, in which a sample experiences repetitive cyclic strain, will stabilize the microstructures of SELP-47K films and improve their resilience under subsequent deformation that is lower than the preconditioning strain. Mechanical Preconditioning Improved the Resilience of SELP-47K Films. The influence of mechanical preconditioning on the material behavior of methanol-treated non-cross-linked films and cross-linked films was examined by subjecting hydrated samples to repetitive cyclic displacements of 5 mm, which corresponded to about 45% strain. We observed the accumulation of a small residual deformation and a slight decline



3031

Figure 6. Representative stress-strain curves of mechanically preconditioned films. After 10 cycles of preconditioning at ∼45% strain (Figure 4A,B), both methanol-treated non-cross-linked and crosslinked SELP-47K films demonstrated highly linear elasticity.

Figure 4. Representative mechanical preconditioning of methanoltreated noncross-linked (A) and cross-linked SELP-47K films (B). Both samples were cyclically stretched to a maximum displacement of 5 mm for the first 10 cycles. Repeatable stress-strain curves were obtained in both samples after 5-6 cycles of preconditioning, suggesting their microstructures were stabilized.

Figure 5. Resilience of methanol-treated non-cross-linked (() and cross-linked (b) SELP-47K films (n ) 6). The resilience of non-crosslinked films increased from 66 ( 4 to 86 ( 4%, and that of crosslinked films increased from 77 ( 2 to 88 ( 1% upon 10 cycles of preconditioning. For clarity, only the negative error bars of the resilience of noncross-linked films are plotted. The apparent small drop in resilience at cycle 3 was due to the manual repositioning of the sample to remove the buckling that occurred after two cycles of loading and unloading.

in peak stress that stabilized after several loading cycles in both methanol-treated non-cross-linked (Figure 4A) and cross-linked films (Figure 4B). The small residual strains (around 5%) were largely due to deformation-induced structural changes, although realignment of an imperfectly loaded sample was responsible for small lags observed in the first loading curve. Without preconditioning, cross-linked films possessed an enhanced

resilience (77 ( 2%, Figure 4B) over their methanol-treated non-cross-linked counterparts (resilience of 66 ( 4%, Figure 4A). The difference in resilience suggests that glutaraldehyde cross-linking likely limited the chain rearrangement of hydrated elastinlike blocks, decreasing the hysteresis of SELP-47K films. Nevertheless, over 10 loading cycles, the resilience of the methanol-treated non-cross-linked films increased from 66 ( 4 to 86 ( 4% (Figure 5 and Table 1, n ) 6). Likewise, the resilience of cross-linked films increased from 77 ( 2 to 88 ( 1% after 10 cycles of preconditioning (Figure 5 and Table 1, n ) 6), a more modest increase in resilience than their methanoltreated non-cross-linked counterparts. The greatest increase in resilience largely occurred after the first cycle, presumably due to stabilization of deformation-induced changes in microstructure. It bears significance that the obtained resilience of 86 ( 4% for preconditioned methanol-treated non-cross-linked films and of 88 ( 1% for preconditioned cross-linked films closely matches that of native elastin, which is 90%.22,25 This similarity suggests that under small to medium strain the elastinlike blocks of SELP-47K in mechanically preconditioned films may be solely responsible for deformation while the crystallized silklike blocks remain intact. Mechanically Preconditioned SELP-47K Films Demonstrated Highly Linear Elasticity. Both methanol-treated noncross-linked and cross-linked films displayed linear viscoelastic behavior up to 40% strain (Figure 6). The stress-strain analysis further revealed a Young’s modulus of 1.67 ( 0.37 and 3.34 ( 0.26 MPa for preconditioned methanol-treated non-crosslinked and preconditioned cross-linked films, respectively (Table 1). Therefore, GTA-cross-linking doubled the Young’s modulus of SELP-47K films. Because the lysines capable of GTA-crosslinking are only within the elastinlike blocks of SELP-47K, this reinforced our reasoning that under small to medium strain the elastinlike blocks of SELP-47K films are primarily, if not solely, responsible for deformation. Stress Relaxation and Creep Analysis Confirmed the Excellent Elasticity of Preconditioned SELP-47K Films. When deformation was held at 20% strain, the initial stress was 683 KPa in mechanically preconditioned cross-linked films and 313 KPa in preconditioned non-cross-linked films (Figure 7).

Table 1. Comparison of Young’s Modulus (E), Resilience (R), and Tensile Strength at 40% Strain (σ40) of Methanol-Treated Non-Cross-Linked and Cross-Linked SELP-47K Films

non-cross-linked cross-linked

E at 1st cycle (MPa)

E at 10th cycle (MPa)

R at 1st cycle (%)

R at 10th cycle (%)

σ40 (MPa)

1.57 ( 0.3 3.17 ( 0.23

1.66 ( 0.37 3.34 ( 0.26

66 ( 4 77 ( 2

86 ( 4 88 ( 1

0.63 ( 0.14 1.36 ( 0.11

3032


Teng et al.

Figure 7. Representative stress relaxation of preconditioned noncross-linked and cross-linked SELP-47K films at 20% constant strain.

Figure 8. Representative creep of preconditioned cross-linked (solid line) and preconditioned non-cross-linked (dashed line) SELP-47K films at 900 and 400 KPa constant stress, respectively. The stressstrain analysis of preconditioned methanol-treated non-cross-linked and cross-linked SELP-47K films revealed a linear viscoelastic region up to 40% for both types of films. Different stress levels were chosen for creep tests of preconditioned methanol-treated non-cross-linked and cross-linked SELP-47K films to ensure that the total deformation was within 40% strain, in particular, around 30% strain.

Figure 10. Surface morphology of SELP-47K films examined by SEM: nontreated (A); methanol-treated non-cross-linked (B); and methanoltreated and then glutaraldehyde-cross-linked (C).

Figure 9. Methanol treatment induced conformational conversion of SELP-47K films from silk I to silk II structure: Raman spectra of nontreated films (a); methanol-treated non-cross-linked films (b); and methanol-treated and then mechanically preconditioned films (c). The spectra were normalized to the absorbance of the methylene bending band at 1450 cm-1, which is insensitive to changes in secondary structure. Raman marker bands of silk I and silk II are marked by blue and red arrows, respectively.

Stress in preconditioned cross-linked films dropped to 645 KPa in the first 5 min and stabilized thereafter at 620 KPa. Compared to a 9% decrease in stress of cross-linked films prior to equilibrium, non-cross-linked SELP-47K films exhibited a 16% decrease in stress from 313 to 260 KPa in 15 h.

Creep analysis was conducted by placing films under constant stress that would produce an initial approximate strain of 20-30%. Under a sustained stress of 900 KPa, preconditioned cross-linked SELP-47K films exhibited a creep strain of 4%, from an instantaneous deformation of 25.3% to a final deformation of 29.2% after 15 h (Figure 8). Likewise, creep for preconditioned non-cross-linked films under 400 KPa stress for 15 h approached 5%. These modest creep and stress relaxation results suggested that medium deformations (e.g., 20%) and mechanical stresses (e.g., 900 KPa) induced minimal damage to preconditioned SELP-47K films and that their deformation response was elastic in nature. Raman Spectroscopy Revealed the Crystallization of the Silklike Blocks of SELP-47K Films Induced by Methanol Treatment. Nonsolvents such as methanol are often used as a coagulant in artificial silk fiber spinning to remove water and to induce the irreversible formation of β-sheet crystals.28 A Raman analysis of the nontreated films in the 1700-800 cm-1



3033

Figure 11. Surface morphology of methanol-treated non-cross-linked (A) and methanol-treated glutaraldehyde-cross-linked (B) SELP-47K films after mechanical preconditioning. The direction of the preconditioning strain is indicated by the arrows.

spectral region revealed Raman marker bands of silk I structure, including 1248 cm-1 (amide III29), 1102 cm-1 (CC skeletal stretching30 and C(CH3)2 rocking31), 954 cm-1 (CH3 rocking32,33), and 856 cm-1 (CC and CN stretching;33,34 Figure 9). Upon methanol treatment, SELP-47K films displayed Raman marker bands of silk II structure, such as 1230, 1085, 971, and 878 cm-1.29,35 It is well established that silk fibroins (SF) exist in two distinct crystalline forms: silk I and silk II. While silk II is largely comprised of antiparallel β-sheets, the conformation of silk I has not been well understood. Because the sequence analysis of the B. mori fibroin36 revealed that the crystalline regions of silk fibroins contain a highly repetitive GAGAGS sequence, poly(Ala-Gly) copolymers have long been used as a model for the study of the silk I structure.37 Based on solidstate NMR analyses of isotopically labeled poly(Ala-Gly) copolymers, Asakura and colleagues proposed a repeated type II β-turn structure for silk I.38,39 A solid-state NMR analysis further suggested that poly(GAGAGS) has a stronger propensity than poly(Ala-Gly) to form the silk II structure after various treatments including air drying.40 However, our Raman spectral analysis here indicates that air-dried SELP-47K films exist predominantly as silk I. Because SELP-47K has a monomer structure of (S)4(E)4(EK)(E)3, it is likely that the large elastinlike block (E)4(EK)(E)3 prevents the four repeats of the silklike sequence GAGAGS from being fully crystallized into insoluble silk II. A similar disruptive effect on the formation of silk II structure of poly(GAGAGS) was reported when other sequences with hydrophobic side chains (e.g., Y, V) were incorporated.40 Nevertheless, methanol treatment of SELP-47K films induced a conformational conversion from Silk I to insoluble Silk II, which greatly impacted the material stability and mechanical properties of the SELP-47K films.

Figure 12. Fracture surface of methanol-treated non-cross-linked SELP-47K films.

In contrast, mechanical preconditioning did not induce any appreciable changes in the Raman spectra of methanol-treated noncross-linked SELP-47K films. We were not able to assess the possible changes in conformation, if any, after chemical cross-linking because strong fluorescent effects likely induced by the formation of CdN bonds due to glutaraldehyde (GTA) cross-linking prevented the collection of a meaningful Raman spectrum of cross-linked SELP-47K films. It is also worthy of mentioning that the shift of the peaks in the Raman spectrum of SELP-47K after methanol treatment is not due to the effects of any methanol remaining. Mammone et al. reported that methanol displayed very strong anti- and symmetric CH3 stretching bands at 2940 and 2832 cm-1.41 In contrast, the antiand symmetric CH3 stretching bands of large molecules such as SELP-47K appeared at 2972 and 2875 cm-1, which were observed in the Raman spectra of methanol-treated SELP-47K films (see the Raman spectra of SELP-47K films in the spectral range of 400-4000 cm-1, Supporting Information). In addition, the strong C-O stretching band of methanol at 1033 cm-1 did

3034


Teng et al.

Table 2. Comparison of the Elastic Properties of Native Elastin, with Those of Physically- and Chemically-Cross-Linked ELP Polymersa elastic properties elastin type aortic elastin44 elastin (bovine ligament)22,25 exons 20-24-2444 exons 20-(24)444 exons 20-2445 poly(ELP)10

elastomeric unit

(PGVGV/LA)n (PGVGV/LA)n (PGVGV/LA)n

poly(ELP)9

CS5-(VPGIG)20 [CS5-(VPGIG)20]3 [CS5-(VPGIG)20]5 CS5-(VPGIG)25 [CS5-(VPGIG)25]3 [CS5-(VPGIG)25]5 [(VPGXG)5]4c

ELP[KV6]-11211 ELP[KV16]-10211 ELP[KV7F]-7247 ELP[KV2F]-6447 triblock B912

(VPGXG)112d (VPGXG)102g (VPGXG)72h (VPGXG)64i ((VPGXG)5)48j

poly(ELP)46

triblock B10

15

Triblock LysB1014 SELP-47K (this study)

k 5 20

((VPGXG) )

((VPGXG)5)28k (GXGVP)

m 8

cross-linking

PQQ PQQ genipin PQQ film, GTA film, HDMI hydrogel, BS3 casting, DSS hydrogel, TSAT hydrogel, TSAT hydrogel, THPP hydrogel, THPP casting at 5 °C casting at 23 °C casting at 4 °C casting at 23 °C casting at 23 °C casting, GTA coagulant coagulant/GTA

E (MPa)

σT (MPa)

εf (%)

r (%)

0.81 1.1 0.25 ( 0.1 0.25 ( 0.09 1.8 0.4 0.32 ( 0.06 0.2 ( 0.04 0.1 ( 0.03 0.71 ( 0.04 0.58 ( 0.02 0.13 ( 0.03 0.08-0.7 0.3-0.97 0.0016-0.013e 0.0036-0.013e 0.0058-0.0128e 0.0258-0.0458e 1.3 ( 0.3 0.01-0.03 60 ( 8 0.71 ( 0.12 0.49 ( 0.03 1.10 ( 0.45 1.66 ( 0.37 3.34 ( 0.26

1.02 2.0 0.19 ( 0.08 0.23 ( 0.08

103 150 86 ( 42 103 ( 24 68 ( 52 90 ( 37 251 ( 52 393 ( 13 389 ( 14 340 ( 20 480 ( 40 570 ( 30 100-200

77 ( 2 90 80 ( 7 77 ( 10 90.3 ( 6.2b 79.6 ( 1.6b

0.53 ( 0.29 0.87 ( 0.19 0.35 ( 0.06

2.87 ( 0.88 0.78 ( 0.28

640 ( 116 1084 ( 67

2.88 ( 0.91 3.62 ( 0.98 2.5 ( 0.4 5.4 ( 1.1

430 ( 34 223 ( 30 190 ( 60 245 ( 58

δf: 1-4° δ: 4-16° δ: 1-3° δ: 1-3° 39 ( 2%l 51 ( 2%l 52 ( 2 39 ( 1 86 ( 4 88 ( 1

a All polymers were tested in hydrated state. Abbreviations: PQQ, pyrroloquinoline quinine; CS5, REDV containing cell binding domain GEEIQIGHIPREDVDYHYP; GTA, glutaraldehyde; HMDI, hexamethylene diisocyanate; BS3, bis(sulfosuccinimidyl) suberate; DSS, disuccinimidyl suberate; TSAT, tris-succinimidyl aminotriacetrate; THPP, tris(hydroxymethyl)phosphino-propionic acid; E, elastic modulus; σf, ultimate tensile strength; εf, strain at failure; r, resilience. b Resilience was calculated from the energy loss provided in reference: resilience ) (1 - energy loss) × 100%. c X was chosen to be lysine every one of five pentapeptides; otherwise, X was isoleucine. d X was chosen to be lysine every one of seven pentapeptides; otherwise, X was valine. e Complex modulus. f Loss angle (δ). g X was chosen to be lysine every one of 17 pentapeptides; otherwise, X was valine. h X was chosen to be lysine every one of nine pentapeptides and to be phenylalanine every one of nine pentapeptides; otherwise, X was valine. i X was chosen to be lysine every one of four pentapeptides and to be phenylalanine every one of four pentapeptides; otherwise, X was valine. j X was chosen to be glutamic acid every one of five pentapeptides; otherwise, X was valine. k X was chosen to be glutamic acid every one of five pentapeptides; otherwise, X was alanine. l Resilience measured at 30% cyclic strains. m X was chosen to be lysine every one of eight pentapeptides; otherwise, X was valine.

not show in the Raman spectra of SELP-47K films. Thus, it is crystal structural change, not methanol remaining, that induced the shift of the Raman spectra of SELP-47K films. SEM Analysis Revealed Changes in the Surface Microstructures of SELP-47K Films Induced by Methanol Treatment and Mechanical Preconditioning. Along with the conformational conversion revealed by Raman spectroscopy, methanol treatment also induced changes in the surface morphology and microstructure of SELP-47K films. SEM analysis revealed a microfibrilar structure of nontreated SELP-47K films as prepared from aqueous solution (Figure 10A). A close examination of the SEM images suggested that these randomly oriented microfibrils were formed by many protein globules of micrometers in size. This observation is consistent with the hierarchical self-assembling model proposed by Jin and Kaplan for silk.42 According to their model, silk proteins in aqueous solutions form micelles of 100-200 nm in size and these micelles further aggregate into larger globules due to hydrophobic folding and hydrophilic interaction. In SELP-47K aqueous solution, the elastinlike blocks GVGVP may form a relatively loose and lesshydrated core, while the more hydrophilic silklike blocks GAGAGS, the pentapeptide sequence GKGVP with a charged lysine residue, and the two nonrepetitive amino acid sequences at the N- and C-termini of the polymer containing more charged residues may form a corona. The core-corona micelles form large globules due to weak hydrophilic interaction. Consequently, nontreated SELP-47K films were unstable under mechanical forces when fully hydrated. Upon methanol treatment, the silklike blocks of SELP-47K molecules coagulated into insoluble silk II crystals, stabilized by strong intermolecular

hydrogen bonding. The insoluble β-sheet crystals likely formed an interconnected network, linked by the flexible elastinlike blocks. As a result, SEM analysis of SELP-47K films revealed that the aggregates of core-corona micelles disappeared after methanol treatment (Figure 10B). Instead, the methanol-treated non-cross-linked SELP-47K films had a wrinkled surface likely due to the coagulation effects of the methanol treatment. SEM analysis further revealed that GTA-cross-linking of methanoltreated SELP-47K films induced some changes, although subtle, in their microstructure and surface morphology (Figure 10C). While mechanical preconditioning is often used to stabilize the microstructures of soft materials in order to obtain more consistent mechanical properties, the underlying conformational and microstructural changes induced by this processing have rarely been evaluated.43 We examined mechanically preconditioned SELP-47K films by SEM to determine if microstructural changes would correlate with the mechanical changes observed. Methanol-treated non-cross-linked SELP-47K films revealed the deformation-induced, anisotropic structural changes of mechanical preconditioning. In contrast to the isotropic surface morphology of SELP-47K films (Figure 10B), a fiber-like texture appeared on the film surface after mechanical preconditioning and the texture was aligned along the direction of the applied strain (Figure 11A). It is likely that mechanical preconditioning induced molecular alignment and led to the formation of the anisotropic microstructures. Interestingly, less anisotropic surface morphology was observed for cross-linked SELP-47K films after mechanical preconditioning (Figure 11B). In this case, GTAcross-linking likely reduced the molecular alignment and


enhancement of intermolecular hydrogen bonding possible between molecules along the direction of the preconditioning strain. SEM Analysis Suggested Ductile Fracture as a Failure Mode for SELP-47K Films. The fracture surfaces of methanoltreated non-cross-linked SELP-47K films were also analyzed by SEM (Figure 12A-C). Consistent with the high deformability revealed by the tensile analysis, the rough fracture surfaces suggest ductile facture as a failure mode for the fully hydrated SELP-47K films (Figure 12B,C). A close examination of the SEM images suggests that a crack was likely initiated in region B (Figure 12A). The accumulated strain energy may have induced the disruption of the crystallized physical cross-links and the pulling-out of polymer chains and hard phases, resulting in the formation of a rough, dimpled fracture surface (Figure 12B). As the crack propagated and the strain energy was gradually released, the pulling-out of polymer chains and hard phases became less feasible and, as a consequence, relatively smoother fracture surfaces were observed (Figure 12C). Comparison of the Mechanical Properties of SELP-47K Films with Those of Other Chemically and Physically CrossLinked ELP Polymers. A number of strategies have been explored in the development of ELP polymers with properties mimicking those of native elastin. Notably, recombinant ELPs based on human tropoelastin displayed mechanical properties (e.g., elastic modulus, strain at failure, resilience) that closely matched or even exceeded those of native aortic elastin (Table 2).44,45 However, the tropoelastin-derived ELP polymers need to be cross-linked for tissue engineering applications using potentially cytotoxic reagents. Pentapeptide sequences VPGXG, in which X is a guest residue, are also used in the design of ELP polymers.9-11,46,47 Rheological studies revealed small loss angles of ELP hydrogels, suggesting excellent resilience at small strains (Table 2).11,47 However, the pentapeptide-based ELP polymers often possessed much lower elastic modulus than native elastin. Like the tropoelastin-derived ELP polymers, the VPGXG-based ELPs need to be chemically cross-linked to obtain sufficient mechanical stability. The in vivo cross-linking of native elastin is achieved through the enzymatic oxidation of lysine. Likewise, the chemical crosslinking of ELP materials can be obtained through ligation of lysine residues from adjacent molecules using various reagents. The use of such reagents as tris-succinimidyl aminotriacetrate (TSAT) and tris(hydroxymethyl)phosphino-propionic acid (THPP), as reported by the Chilkoti group, resulted in an extremely low elastic modulus of the chemically cross-linked ELP hydrogels.47 Strategies employing the use of physical cross-linking mechanisms to produce sufficient mechanical stability were explored in the synthesis of triblock ELP polymers.12,14,15 Nevertheless, studies revealed severe plastic deformation and poor resilience of those physically cross-linked triblock ELP polymers.12,15 The resilience of physically cross-linked triblock ELPs was enhanced by further chemical cross-linking, but it was still inferior compared to that of native elastin.14 In contrast, we demonstrated herein that SELP-47K stabilized only by physical cross-links formed by the crystallization of the silklike blocks displayed an elastic modulus, ultimate tensile strength, extensibility, and resilience, comparable to or exceeding those properties of native aortic elastin44 and elastin fibers from bovine ligaments.22,25 Further, chemical cross-linking doubled the elastic modulus and tensile strength of SELP-47K films and enhanced their extensibility and resilience. Thus, the incorporation of crystallizable silklike blocks in SELPs appears


3035

to be an attractive alternative strategy in the design of elastin-mimetic materials such as SELPs.

Conclusion Recombinant silk-elastinlike protein copolymer SELP-47K films after mechanical preconditioning displayed excellent elasticity, evident by tensile stress-strain, creep, and stress relaxation analyses. In particular, the mechanical properties of methanol-treated non-cross-linked SELP-47K films, including Young’s modulus, resilience, deformability, and tensile strength, closely match those of native aortic elastin and elastin fibers from bovine ligaments. The Young’s modulus of SELP-47K films was further doubled by glutaraldehyde cross-linking. A Raman spectroscopic analysis revealed the crystallization of the silklike sequences of SELP-47K films upon methanol treatment, leading to the formation of insoluble physical cross-links. Our results suggest that under small to medium strains the elastinlike blocks of SELP-47K alone deform, while the physical crosslinks formed by the crystallized silklike blocks remain intact, explaining the excellent elasticity and resilience of SELP-47K films. Given its mechanical properties and its processability, SELP-47K may find a wide range of biomedical and tissue engineering applications as a suitable substitute for elastin. Acknowledgment. We thank Professor Robert Downs for assistance in collecting the Raman spectra of SELP-47K films. This work was funded by an NSF grant (CMMI 0700323). Supporting Information Available. Raman spectrum of methanol-treated SELP-47K films. This material is available free of charge via the Internet at http://pubs.acs.org.

References and Notes (1) Vrhovski, B.; Weiss, A. S. Eur. J. Biochem. 1998, 258 (1), 1–18. (2) Uitto, J.; Christiano, A. M.; Kahari, V. M.; Bashir, M. M.; Rosenbloom, J. Biochem. Soc. Trans. 1991, 19 (4), 824–829. (3) Privitera, S.; Prody, C. A.; Callahan, J. W.; Hinek, A. J. Biol. Chem. 1998, 273 (11), 6319–6326. (4) Urry, D. W. Perspect. Biol. Med. 1978, 21 (2), 265–295. (5) Vrhovski, B.; Jensen, S.; Weiss, A. S. Eur. J. Biochem. 1997, 250 (1), 92–98. (6) Debelle, L.; Alix, A. J. P.; Jacob, M. P.; Huvenne, J. P.; Berjot, M.; Sombret, B.; Legrand, P. J. Biol. Chem. 1995, 270 (44), 26099–26103. (7) Tamburro, A. M.; Guantieri, V.; Dagagordini, D.; Abatangelo, G. Biochim. Biophys. Acta 1977, 492 (2), 370–376. (8) Li, D. Y.; Brooke, B.; Davis, E. C.; Mecham, R. P.; Sorensen, L. K.; Boak, B. B.; Eichwald, E.; Keating, M. T. Nature 1998, 393 (6682), 276–280. (9) Di Zio, K.; Tirrell, D. A. Macromolecules 2003, 36 (5), 1553–1558. (10) Welsh, E. R.; Tirrell, D. A. Biomacromolecules 2000, 1 (1), 23–30. (11) Trabbic-Carlson, K.; Setton, L. A.; Chilkoti, A. Biomacromolecules 2003, 4 (3), 572–580. (12) Nagapudi, K.; Brinkman, W. T.; Leisen, J.; Thomas, B. S.; Wright, E. R.; Haller, C.; Wu, X. Y.; Apkarian, R. P.; Conticello, V. P.; Chaikof, E. L. Macromolecules 2005, 38 (2), 345–354. (13) Wright, E. R.; McMillan, R. A.; Cooper, A.; Apkarian, R. P.; Conticello, V. P. AdV. Funct. Mater. 2002, 12 (2), 149–154. (14) Sallach, R. E.; Cui, W.; Wen, J.; Martinez, A.; Conticello, V. P.; Chaikof, E. L. Biomaterials 2009, 30 (3), 409–422. (15) Wu, X. Y.; Sallach, R.; Caves, J.; Conticello, V. P.; Chaikof, E. L. Biomacromolecules 2008, (9), 1787–1794. (16) Wu, X. Y.; Sallach, R.; Haller, C. A.; Caves, J. A.; Nagapudi, K.; Conticello, V. P.; Levenston, M. E.; Chaikof, E. L. Biomacromolecules 2005, 6 (6), 3037–3044. (17) Cappello, J.; Crissman, J. W.; Crissman, M.; Ferrari, F. A.; Textor, G.; Wallis, O.; Whitledge, J. R.; Zhou, X.; Burman, D.; Aukerman, L.; Stedronsky, E. R. J. Controlled Release 1998, 53 (1-3), 105– 117. (18) Ferrari, F.; Cappello, J. In Protein-Based Materials; McGrath, K. P., Kaplan, D. L., Eds.; Birkha¨user: Boston, 1997; pp 37-117.

3036


(19) Keyes, J.; Junkin, M.; Cappello, J.; Wu, X. Y.; Wong, P. K. Appl. Phys. Lett. 2008, (93), 023120. (20) Qiu, W. G.; Teng, W. B.; Cappello, J.; Wu, X. Y. Biomacromolecules 2009, 10 (3), 602–608. (21) Bigi, A.; Cojazzi, G.; Panzavolta, S.; Rubini, K.; Roveri, N. Biomaterials 2001, 22 (8), 763–768. (22) Gosline, J.; Lillie, M.; Carrington, E.; Guerette, P.; Ortlepp, C.; Savage, K. Philos. Trans. R. Soc., B 2002, 357 (1418), 121–132. (23) Gosline, J. M.; Guerette, P. A.; Ortlepp, C. S.; Savage, K. N. J. Exp. Biol. 1999, 202 (23), 3295–3303. (24) Shao, Z. Z.; Vollrath, F. Polymer 1999, 40 (7), 1799–1806. (25) Aaron, B. B.; Gosline, J. M. Biopolymers 1981, 20 (6), 1247–1260. (26) Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953; p 672. (27) Gennes, P.-G. d. Scaling Concepts in Polymer Physics; Cornell University Press: Ithaca, NY, 1979; p 324. (28) Martin, D. C.; Jiang, T.; Buchko, C. J. In Protein-Based Materials; McGrath, K. P., Kaplan, D. L., Eds.; Birkha¨user: Boston, 1997; pp 339-370. (29) Monti, P.; Taddei, P.; Freddi, G.; Asakura, T.; Tsukada, M. J. Raman Spectrosc. 2001, 32 (2), 103–107. (30) Magoshi, J.; Magoshi, Y.; Nakamura, S. Polym. Commun. 1985, 26 (10), 309–311. (31) Koenig, J. L. J. Polym. Sci.: Macromol. ReV. 1972, 6 (1), 59–177. (32) Edwards, H. G. M.; Farwell, D. W. J. Raman Spectrosc. 1995, 26 (8-9), 901–909. (33) Taddei, P.; Asakura, T.; Yao, J. M.; Monti, P. Biopolymers 2004, 75 (4), 314–324. (34) Moore, W. H.; Krimm, S. Biopolymers 1976, 15 (12), 2465–2483.

Teng et al. (35) Zheng, S. D.; Li, G. X.; Yao, W. H.; Yu, T. Y. Appl. Spectrosc. 1989, 43 (7), 1269–1272. (36) Zhou, C. Z.; Confalonieri, F.; Medina, N.; Zivanovic, Y.; Esnault, C.; Yang, T.; Jacquet, M.; Janin, J.; Duguet, M.; Perasso, R.; Li, Z. G. Nucleic Acids Res. 2000, 28 (12), 2413–2419. (37) Lotz, B.; Keith, H. D. J. Mol. Biol. 1971, 61, 1, 201. (38) Asakura, T.; Ashida, J.; Yamane, T.; Kameda, T.; Nakazawa, Y.; Ohgo, K.; Komatsu, K. J. Mol. Biol. 2001, 306 (2), 291–305. (39) Asakura, T.; Yao, J. M.; Yamane, T.; Umemura, K.; Ultrich, A. S. J. Am. Chem. Soc. 2002, 124 (30), 8794–8795. (40) Asakura, T.; Ohgo, K.; Ishida, T.; Taddei, P.; Monti, P.; Kishore, R. Biomacromolecules 2005, 6 (1), 468–474. (41) Mammone, J. F.; Sharma, S. K.; Nicol, M. J. Phys. Chem. 1980, 84 (23), 3130–3134. (42) Jin, H. J.; Kaplan, D. L. Nature 2003, 424 (6952), 1057–1061. (43) Fung, Y. C. Biomechanics: Mechanical Properties of LiVing Tissues, 2nd ed.; Springer-Verlag: New York, 1993; p 568. (44) Bellingham, C. M.; Lillie, M. A.; Gosline, J. M.; Wright, G. M.; Starcher, B. C.; Bailey, A. J.; Woodhouse, K. A.; Keeley, F. W. Biopolymers 2003, 70 (4), 445–455. (45) Vieth, S.; Bellingham, C. M.; Keeley, E. W.; Hodge, S. M.; Rousseau, D. Biopolymers 2007, 85 (3), 199–206. (46) Nowatzki, P. J.; Tirrell, D. A. Biomaterials 2004, 25 (7-8), 1261– 1267. (47) Lim, D. W.; Nettles, D. L.; Setton, L. A.; Chilkoti, A. Biomacromolecules 2007, 8 (5), 1463–1470.

BM900651G

Recombinant Silk-Elastinlike Protein Polymer Displays Elasticity

Recommend Documents